Innovating Online Ethics Training Using a Dynamic Narrative Learning Environment

伦理困境出现在STEM研究的各个方面。为了帮助减轻与不道德的研究行为相关的问题，大学和资助机构，如NSF，需要在负责任的研究行为（RCR）的培训。然而，虽然有各种各样的方法来教授道德行为，包括案例研究，研讨会和工具包，道德培训教学往往减少到简化的基于文本的在线教程。道德监督机构和培训专家强调，需要更多的RCR学习工具，这些工具更具吸引力，可扩展性，并提供探索真实的世界问题的实践方法。该项目探讨了在基于教育游戏的道德培训工具中使用人工智能（AI）技术，以最大限度地提高玩家的动机和学习。具体来说，人工智能将用于创建现实的交互式道德场景，以适应玩家和他们做出的选择。这项研究将提供一种新的基于人工智能的方法来学习道德，并制作一种基于游戏的道德教育培训工具，该工具可以免费访问，并随时准备纳入全国范围内负责任的研究培训计划。此外，通过加强用于培训科学家理解和参与道德行为的方法，本研究旨在推动STEM研究质量和准确性的提高，并鼓励支持多元化和包容性环境的更健康的工作场所实践。和实践道德培训，通过应用自我决定理论（SDT）的动机，通过四个国家的最先进的基于人工智能的叙事系统的创新组合：1）社交模拟系统，2）教学管理器，3）戏剧管理器，以及4）动态内容选择架构（CSA），其基于由前三个系统维护的作者目标和状态来动态地选择文本和选项。与现有的动态CSA方法相比，这种方法能够更灵活地交付基于选择的故事内容的交互式故事。该项目在模拟复杂的伦理背景和促进学习者的自主性，相关性和能力的感觉方面做出了贡献。这些贡献的形式是人工智能驱动的叙事技术，以最大限度地提高不同的学习者动机，一种新颖的动态CSA架构，以及一种利用人工智能来提高学习者参与度，知识和道德推理技能发展的道德培训新方法。该项目采用基于设计的研究方法，迭代构建每个基于AI的叙事系统，并评估这种新的AI叙事训练架构对玩家动机，参与度和道德培训学习成果的影响。还将采用随机对照试验来比较具有动态CSA的自适应交互式叙事游戏与相同游戏的静态叙事版本的功效。研究结果有望促进对基于人工智能的方法的理解，从而在教育叙事和道德培训中利用SDT，这将使学习技术进一步创新。该奖项反映了NSF的法定使命，并被认为值得通过使用基金会的智力价值和更广泛的影响审查标准进行评估来支持。

Ethical dilemmas arise in all aspects of STEM research. To help mitigate the problems associated with unethical research behavior, universities and granting agencies, such as the NSF, require training in responsible conduct of research (RCR). However, while there are a variety of approaches to teaching ethical behavior, including case studies, workshops, and toolkits, ethics training instruction is often reduced to simplified text-based online tutorials. Ethical oversight agencies and training experts have emphasized the need for additional RCR learning tools that are more engaging, scalable, and offer a hands-on approach to exploring real world problems. This project explores the use of artificial intelligence (AI) techniques within an educational game-based ethics training tool to maximize player motivation and learning. Specifically, the AI will be used to create realistic interactive ethical scenarios that adjust to the player and choices they make. This research will provide a novel artificial intelligence-based approach to learning ethics and produce an educational game-based ethics training tool that is freely accessible and ready for incorporation into responsible conduct of research training programs nationwide. Furthermore, by enhancing methods used for training scientists to understand and engage in ethical behavior, this research aims to drive improvements in the quality and accuracy of STEM research as well as encourage healthier workplace practices that support a diverse and inclusive environment.This project will address the need for engaging, scalable, and hands-on ethics training by applying the Self-Determination Theory (SDT) of motivation through an innovative combination of four state-of-the-art AI-based narrative systems: 1) a social simulation system, 2) a pedagogy manager, 3) a drama manager, and 4) a dynamic content selection architecture (CSA) that dynamically selects texts and options based on author goals and state maintained by the first three systems. Compared to existing approaches to dynamic CSA, this approach enables interactive stories with more flexible delivery of choice-based story content. The project develops contributions in modeling complex ethical contexts and facilitating learners’ feelings of autonomy, relatedness, and competence. The forms of these contributions are AI-driven narrative techniques to maximize diverse learner motivations, a novel dynamic CSA architecture, and a new approach to ethics training that utilizes AI to improve learner engagement, knowledge, and moral reasoning skill development. The project takes a design-based research approach to iteratively build out each AI-based narrative system and evaluates the impact this new AI narrative training architecture has on player motivation, engagement, and learning outcomes for ethics training. A randomized controlled trial will also be employed to compare the efficacy of the adaptive interactive narrative game with dynamic CSA to a static narrative version of the same game. The results are expected to advance understanding of AI-based approaches to leveraging SDT within educational narratives and ethics training, which will enable further innovations in learning technologies.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.